Mitochondrial Membrane Potential and Hydroethidine-Monitored Superoxide Generation in Cultured Cerebellar Granule Cells Samantha L

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

FEBS 19236 FEBS Letters 415 (1997) 21-24

Mitochondrial membrane potential and hydroethidine-monitored Superoxide generation in cultured cerebellar granule cells Samantha L. Budd, Roger F. Castilho, David G. Nicholls* Neurosciences Institute, Department of Pharmacology, University of Dundee, Dundee DDl 9S Y, UK Received 18 July 1997; revised version received 6 August 1997

the latter is abolished by protonophores or even the modest Abstract Mitochondrial depolarisation has been reported to enhance the generation of Superoxide anión (Ο'Γ) in a number of decrease in A\|/m during ATP synthesis (reviewed in [1]). cell preparations while an inhibition has been observed with There is considerable current interest in the role of mito- isolated mitochondria. Cerebellar granule cells equilibrated with chondrial free radicals in the initiation of cellular necrosis and > 1 uM hydroethidine (dihy droethidiuni) which is oxidised to the apoptosis in the brain and immune system (reviewed in [6-8]). fluorescent ethidium cation by 0*2~ showed a large increase in However, there is ambiguity in studies with isolated cells as to fluorescence on protonophore addition. However, controls whether protonophores inhibit [9] or, in contradiction to iso- showed the fluorescent enhancement to be a consequence of lated mitochondrial studies, enhance [10-15] the production of release of unbound preformed ethidium from the mitochondrial reactive oxygen species. Many of these latter studies have matrix within the cell with resultant fluorescent enhancement. assayed 0 ~ production by monitoring the oxidation of hy- This ambiguity was removed by the use of low (1 μΜ) 2 droethidine (HET, also called dihydroethidium) to the fluo- concentrations of dye in which case generated ethidium remained bound within the mitochondria. Under these conditions antimycin rescent ethidium [11-15] and in this paper we attempt to re- A, but not protonophore addition, produced an increase in solve the nature of this discrepancy by comparing the fluorescence. It is concluded that excess ethidium acts as an behaviour of this indicator system in both intact cells (cul- indicator of mitochondrial membrane potential obscuring the tured cerebellar granule cells) and isolated rat liver mitochon- monitoring of 0'2~ and that certain experiments employing this dria. Our conclusion is that while HET is an authentic detec- indicator in cells may require re-evaluation. tion system for 02~, the mitochondrial membrane-potential- © 1997 Federation of European Biochemical Societies. dependent distribution of excess, unbound ethidium cation across the mitochondrial inner membrane critically affects its Key words: Mitochondrion; Reactive oxygen species; fluorescent yield. As first reported by Rottenberg [16], ethidi- Neuron; Mitochondrial membrane potential; Superoxide um distributes across the mitochondrial membrane in response to the membrane potential, A\|/m: thus intracellular mitochondrial depolarisation results in a rapid efflux from 1. Introduction the matrix, allowing the cation to bind to nuclear DNA with an extensive fluorescent enhancement. It is concluded that the enhanced fluorescence, seen when HET-loaded cells A major source of Superoxide (0' ~) in cells is complex III 2 are depolarised, reflects an enhanced fluorescent yield of pre- of the mitochondrial respiratory chain where transient qui- formed ethidium rather than an increase in 0 ~ generation. none radicals are intermediates in the chemiosmotic Q-cycle 2 This ambiguity is avoided by the use of low concentrations of (reviewed in [1]). The large pool of free ubiquinol/ubiquinone HET, in which case an enhanced production of 0 ~ by anti- (UQH2/UQ) which exists in the membrane shuttles electrons 2 mycin A, but not by FCCP, can be observed. to complex III, where oxidation of the quinol takes place in two 1 e~ stages: the first electron is transferred to the Rieske protein with the release of two protons and generation of the semiquinone UQ'~; the second electron is transferred to the 2. Materials and methods cytochrome ¿L haem close to the cytoplasmic face of the membrane. Reoxidation of ¿L involves transfer of the electron Cerebellar granule cells were prepared essentially as described by Courtney et al. [17], from 7-day-old Wistar rats, and plated on poly-D- across the membrane to the second haem (bu) close to the coated coverslips at a density of 280 000 per coverslip and cultured in matrix face. This electron transfer is opposed by the mem- minimal essential medium supplemented with foetal calf serum (10% brane potential (A\|/m) and the steady-state concentration v/v), 25 mM KC1, 30 mM glucose, 2 mM glutamine, 50 U/ml of and half-life of UQ' increases in respiratory state 4 [2,3], penicillin and 50 μg/ml of streptomycin. At 24 h in vitro 10 μΜ enhancing ale" non-enzymatic transfer to molecular oxygen cytosine arabinoside was added. Cells were maintained at 37°C in a humidified atmosphere of 5% CC>2/95% air for 7-9 days. Coverslip- with the generation of 0'2~. Consistent with this, direct deter- mounted cells were incubated with 5 μΜ hydroethidine or 5 μΜ minations with isolated mitochondria of 02~ [1,4] and the ethidium (Figs. 1 and 2) in a medium composed of: 153 mM NaCl, Superoxide dismutase product H2O2 [5] show a steep correla- 3.5 mM KC1, 0.4 mM KH2P04, 20 mM TES, 5 mM NaHCOs, 1.2 tion between A\|/ and the rate of 0 ~ generation, such that mM NaS04, 1.2 mM MgCl2, 1.3 mM CaCl2 and 15 mM glucose for m 2 10 min in the dark at 37°C and rinsed before being visualised at 485 nm excitation and > 515 nm emission with an inverted Nikon Dia- phot epifluorescence microscope equipped with a Life Science Resour- *Corresponding author. Fax: (44) (1382) 667120. ces MiraCal imaging system. In a separate experiment (Fig. 4) 1 μΜ E-mail: [email protected] or 10 μΜ HET was present throughout the experiment. Ethidium fluorescence was relatively photostable under normal illumination Abbreviations: HET, hydroethidine; Av|/m, mitochondrial membrane conditions and auto-oxidation in granule cells was negligible for at potential least 30 min.

3. Results

Depolarisation of mitochondria in situ within cultured cerebellar granule cells by the combination of rotenone to inhibit mitochondrial complex I and oligomycin to inhibit the ATP synthase [18] has been shown to confer a high degree of pro- tection against glutamate excitotoxicity [19]. 02~ is believed to play a central role in this degenerative process [6-8] and experiments with isolated mitochondria (reviewed in [1]) would predict that the depolarisation would inhibit the mitochondrial production of 0'2~. The oxidation of HET to ethidium has been widely employed to monitor 02~ generation within cells and to support a paradoxical enhancement of 02~ gen- Έ eration by factors which depolarise the mitochondrion [11- Z) 15]. This counter-intuitive enhancement of fluorescence was found in the present preparation (Fig. 1A); however, a num- CO ber of features suggest that the fluorescence is primarily monitoring Δψπ! rather than 0'2~: thus oligomycin, which hyper- polarises in situ mitochondria [20], decreased the fluorescence; rotenone plus oligomycin, which causes total depolarisation Ü [18,20], caused a large enhancement, while rotenone alone, O which allows Δψπι to be supported by hydrolysis of cytoplas- to mic ATP and only slightly depolarises the mitochondria [18,20], caused a slight enhancement. ω o1— Since these findings suggest that a generally accepted corre- 3 lation may require re-evaluation, a detailed study was performed. When cultured cerebellar granule cells were incubated with HET for 10 min and viewed by digital imaging, morpho- logically intact cells showed a low fluorescence which in- LU creased only slowly with time (Fig. 1A). As reported by Bin- dokas et al. [13] this fluorescence was punctate (Fig. 2A) and localised to mitochondria which in granule cells occupy an annulus around the large central nucleus. Addition of the protonophore FCCP to cells which have been incubated with HET produced a rapid enhancement of fluorescence (Fig. 1A) and at the same time the distribution of the fluorescence changed from a punctate appearance to a diffuse cytoplasmic/nuclear staining (Fig. 2B). The enhanced signal has been observed in cultured hippo- campal neurones [13] and lymphocytes [11] and has been interpreted as an increased rate of Superoxide generation follow- ing mitochondrial depolarisation. However, the cationic ethidium distributes freely across mitochondrial inner mem- Time min branes in response to the membrane potential [16]; thus an Fig. 1. A, B: Single cell fluorescence of a culture of cerebellar gran- alternative explanation could be that the fluorescent yield of ule cells incubated in the presence of 5 μΜ HET for 20 min. In A, the ethidium produced by the oxidation of HET and accumu- each trace represents the mean fluorescencewithi n the somata of 10 individual cells. In B individual cell responses are shown. Additions lated within the mitochondrion could be enhanced on release were made where indicated of 10 μΜ oligomycin, 5 μΜ rotenone, from the mitochondrial matrix. In order to assess the contri- 100 μΜ xanthine plus 10 mU xanthine oxidase (X/XO), or 2 μΜ bution such a redistribution would make to the signal, a par- FCCP. C: Cells were incubated in the presence of 5 μΜ ethidium. allel experiment was performed in which cells were incubated Note that the fluorescence increase parallels that shown in B. Re- in the presence of 5 μΜ ethidium (Fig. IB). Sufficient ethidi- sults for each experiment are representative of 3^1 independent experiments from different cultures. um accumulated to produce a punctate fluorescence in mor- phologically healthy cells (Fig. 2E); however, addition of FCCP caused a rapid enhancement (Fig. 1A) and diffusion the assay does not detect generated H2C>2, while the assay (Fig. 2F) of the fluorescence which was closely identical to only slowly responds to high concentrations of added H2C>2. that observed in the HET experiment. Thus most or all of Ethidium binds avidly to nucleic acids and Fig. 3B shows the protonophore-induced fluorescence increase in the HET the large fluorescent enhancement produced by the addition of experiment can be ascribed to redistribution of pre-existing DNA to a cell-free assay. For HET to accurately monitor 02~ ethidium rather than enhanced 0'2~ generation. under conditions where A\\im may vary it is therefore necessary HET detects 0'2~ in a cell-free assay (Fig. 3A): the detection to eliminate changes in fluorescent yield due to redistribution of the free radical generated by xanthine/xanthine oxidase is of excess unbound ethidium. This can be achieved by decreas- prevented by addition of Superoxide dismutase, indicating that ing the concentration of HET such that ethidium is not gen- S.L. Budd et al.lFEBS Letters 415 (1997) 21-24 23

Fig. 2. Fluorescence of cerebellar granule cells incubated with 5 μΜ HET or ethidium in response to protonophore-induced depolarization of mitochondria. A: Bright-field digital image of cells prior to loading with 5 μΜ HET. B: Digital fluorescent image of the same field of cells after loading with 5 μΜ HET and incubation for 20 min. C: The same field 2 min after addition of 2 μΜ FCCP. Note the punctate appearance of the fluorescence in B and the enhanced, diffuse fluorescence after addition of the protonophore. D, E, F: Respective bright-field and fluorescent images of cells loaded with 5 μΜ ethidium and imaged before and 2 min after addition of 2 μΜ FCCP. Note that a morphologi- cally damaged cell (arrowed) displays a high fluorescence both before and after FCCP addition.

erated beyond the binding capacity of mitochondrial nucleic acids. Fig. 4 shows that the protonophore-mediated enhancement in fluorescence is only observed in cells equilibrated with high concentrations of HET. The ability to monitor authentic Superoxide generation after equilibration with low concentrations of HET under these conditions is confirmed by the en- +10mM H202 hanced rate of fluorescence increase on addition of antimycin A, consistent with results obtained with isolated mitochondria (reviewed in [1]). + SOD 4. Discussion control

These results help to resolve some of the apparent contra- 0 4 8 12 16 20 c Time (min) dictions between the effect of Ai|/m on 02~ generation by iso- o lated mitochondria and by mitochondria in situ within the 300 functioning cells, and emphasise the complexities inherent in o assaying mitochondrial 0"2~ generation in intact cells. None of £ the currently employed fluorescent probes of reactive oxygen rs 225- is without complications: dichlorodihydrofluorescein deriva- !c LU 150

Fig. 3. A: Oxidation of HET by 02~ generated by xanthine/xanthine oxidase (X/XO) in a cell-free assay: note that HET is oxidized by X/XO but that this is abolished in the presence of 50 U/ml 75· Superoxide dismutase (SOD) indicating that the dye is not oxidized by generated H202. However, high concentrations of added H202 cause a slow increase [14]. Traces are displaced vertically for clarity. B: Fluorescent enhancement of ethidium standards in the presence 5 10 15 of 100 μg/ml DNA. ethidium (nmol) 24 S.L. Budd et al.lFEBS Letters 415 (1997) 21-24

S" 140 140 Έ i 105 105 oΦ c a> υ 70 (0 70 o ^ o _3 35- 35 **- E 3 Έ ιθ Ξ 0 10 20 30 40 50 0 10 20 30 40 50

Time (min)

Fig. 4. Low concentrations of HET allow observation of Oj~ production independent of a A\|/m-dependent artifact. A: Addition of protonophore (1 μΜ FCCP) to cells incubated in the continuous presence of 10 μΜ HET results in a large enhancement of fluorescence. B: Cells incubated in parallel with 1 μΜ HET are insensitive to protonophore addition but show an increased rate of O'j production in response to 1 μΜ antimycin A (AA). In both A and B 1 μg/ml oligomycin is also present throughout to ensure a complete reduction in A\|/m. tives react with H2O2 to give oxidised products which are pH [3: Van Beizen, R., Kotlyar, A.B., Moon, N., Dunham, W.R. and Albracht, S.P.J. (1997) Biochemistry 36, 886-893. dependent and to varying extents cell permeant [9,21]; despite [4: Liu, S.S. and Huang, J.P. (1996) in: Procedings of the Interna- difficulties in deconvoluting the pH effect using this indicator, tional Symposium on Natural Antioxidants (Moores, D., Ed.) protonophores have been reported to inhibit generation of AOCS Press, Champaign, IL (in press). reactive oxygen species in cortical neurones [13]. Dihydro- [5: Boveris, A. and Chance, B. (1973) Biochem. J. 128, 617-630. rhodamine-123 responds also to H2O2 [21] and peroxynitrite te: Golstein, P. (1997) Science 275, 1081-1082. Petit, P.X., Susin, S.A., Zamzami, N., [22,23] and the product rhodamine-123 is again a mitochon- [7 Mignotte, B. and Kroemer, G. (1996) FEBS Lett. 396, 7-13. drial A\|/ indicator; using this indicator protonophores were m [s: Richter, C, Schweizer, M., Cossarizza, A. and Franceschi, C. reported to double production of reactive oxygen species in (1996) FEBS Lett. 378, 107-110. cortical neurones [10]. In a comprehensive study to establish [9: Reynolds, I.J. and Hastings, T.G. (1995) J. Neurosci. 15, 3318- _ 3327. that HET reacted selectively with 02 in cultured hippocam- pal neurones, Bindokas et al. [13] also reported a protono- [iff Dugan, L.L., Sensi, S.L., Canzoniero, L.M.T., Handran, S.D., Rothman, S.M., Lin, T.S., Goldberg, M.P. and Choi, D.W. phore-induced enhancement in signal which was interpreted as (1995) J. Neurosci. 15, 6377-6388. increased 02~ generation. [11 Zamzami, N., Marchetti, P., Castedo, M., Decaudin, D., Macho, The central goal of many related studies is to establish the A., Hirsch, T., Susin, S.A., Petit, P.X., Mignotte, B. and Kroemer, G. (1995) J. Exp. Med. 182, 367-377. chain of causality between physiological or pathological de- Castedo, M., Macho, A., Zamzami, N., Hirsch, T., Marchetti, P., _ [1? creases in mitochondrial polarisation, 02 generation, the mi- Uriel, J. and Kroemer, G. (1995) Eur. J. Immunol. 25, 3277- tochondrial permeability transition and subsequent apoptotic 3284. and necrotic cell death (reviewed in [1,7]). Two major areas [i3: Bindokas, V.P., Jordan, J., Lee, C.C. and Miller, R.J. (1996) where HET has been employed to report a correlation be- J. Neurosci. 16, 1324-1336. [14] Scanion, J.M., Aizenman, E. and Reynolds, I.J. (1997) Eur. tween a decreased A\|/m and an enhanced 0'2~ generation are J. Pharmacol. 326, 67-74. in dexamethasone-induced lymphocyte apoptosis [7] and neu- Macho, A., Castedo, M., Marchetti, P., Aguilar, J.J., Decaudin, ronal glutamate excitotoxicity [13]. We feel that some of these [is: D., Zamzami, N., Girard, P.M., Uriel, J. and Kroemer, G. (1995) conclusions should be re-examined to deconvolute the abso- Blood 86, 2481-2487. lute production of ethidium from the variable fluorescent en- Rottenberg, H. (1984) J. Membr. Biol. 81, 127-138. [i6: Courtney, M.J., Lambert, J.J. and Nicholls, D.G. (1990) J. Neu- hancement depending on compartmentation of this mitochon- [17 rosci. 10, 3873-3879. drial membrane potential indicator [16]. Budd, S.L. and Nicholls, D.G. (1996) J. Neurochem. 66, 403- [is: 410. Budd, S.L. and Nicholls, D.G. (1996) J. Neurochem. 67, 2281- Acknowledgements: This work was supported by the Wellcome Trust. [i? S.L.B. is supported by a Medical Research Council studentship. 2291. [20" Scott, I.D. and Nicholls, D.G. (1980) Biochem. J. 192, 873-880. [21 Royall, J.A. and Ischiropoulos, H. (1993) Arch. Biochem. Bio- References phys. 302, 348-355. [22 Henderson, L.M. and Chappell, J.B. (1993) Eur. J. Biochem. 217, [1] Skulachev, V.P. (1996) Q. Rev. Biophys. 29, 169-202. 973-980. [2] Turrens, J.F., Alexandre, A. and Lehninger, A.L. (1985) Arch. [23: Kooy, N.W., Royall, J.A., Ischiropoulos, H. and Beckman, J.S. Biochem. Biophys. 237, 408^114. (1994) Free Radical Biol. Med. 16, 149-156.